Akka vs. Spring Reactor personal comparison



Akka vs. Spring Reactor personal comparison:

I used for both a mass of Fake data of 200000 records for each.

Both use the Reactive Streams Specification (Reactive Streams)

Reactive Streams is an initiative to provide a standard for asynchronous flow processing with non-blocking return pressure. This includes efforts focused on runtime environments (JVM and JavaScript) as well as network protocols.

What we want is asynchronous and non-blocking. A group of developers of companies like Netflix, Pivotal, RedHat, etc. met and converged to something called The Reactive Streams Specification.
 
Akka framework

Akka is an Opensource framework.

Akka is implemented from the agent model (originated in 1973 by Carl Hewitt), which provides a model for dealing with concurrent computing.

Akka, the unit of execution is the Actor and his microservices are built as actors. An actor is lightweight and there may be several million actors per GB of heap memory.

The actor is an object that encapsulates state and behavior and communicates exclusively by exchanging messages that are placed in the recipient's mailbox. Below is a diagram that illustrates this.

Simpler Concurrent & Distributed Systems

Actors and Streams let you build systems that scale up, using the resources of a server more efficiently, and out, using multiple servers.

Resilient by Design

Building on the principles of The Reactive Manifesto Akka allows you to write systems that self-heal and stay responsive in the face of failures.

High Performance

Up to 50 million msg/sec on a single machine. Small memory footprint; ~2.5 million actors per GB of heap.

Elastic & Decentralized

Distributed systems without single points of failure. Load balancing and adaptive routing across nodes. Event Sourcing and CQRS with Cluster Sharding. Distributed Data for eventual consistency using CRDTs.

Reactive Streaming Data

Asynchronous non-blocking stream processing with backpressure. Fully async and streaming HTTP server and client provides a great platform for building microservices. Streaming integrations with Alpakka.

Time metric using 200000 records for both:

Akka:

-------------------------------------------------------------------------------
Test set: com.lightbend.akka.sample.AkkaQuickstartTest
-------------------------------------------------------------------------------
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.766 sec


INITIAL TIME: [05/10/2019 10:10:49.708]

FINAL TIME: [05/10/2019 10:11:03.081]

Average 13.5 seconds

Spring Reactor


Like Akka Spring is also an Opensource framework.

Project Reactor is Spring's implementation of The Reactive Specification and it's specifically favored by the Spring Webflux module, although you can use it with other modules like RxJava.
The idea is to operate Asynchronously with Backpressure using Publishers and Subscribers.
Here, we're being introduced to several new concepts! Let's explain them one by one:
  • Publisher - A Publisher is a provider of a potentially unbounded number of elements.
  • Subscriber - A Subscriber listens to that Publisher, asking for new data. Sometimes, it's also referred to as a Consumer.
  • Backpressure - The ability of the Subscriber to let the Publisher how many requests can it handle at the time. So it's the Subscriber that is responsible for the flow of the data, not the Publisher as it just provides the data.
The Reactor Project offers 2 types of publishers. These are considered the main building blocks of Spring Webflux:
  • Flux - is a publisher that produces 0 to N values. It could be unbounded. Operations that return multiple elements use this type.
  • Mono - is a publisher that produces 0 to 1 value. Operations that return a single element use this type.

Time metric using 200000 records for both:

Spring Reactor:

-------------------------------------------------------------------------------
Test set: reactivewebconsumer.reactivewebconsumer.ReactiveWebConsumerApplicationTests
-------------------------------------------------------------------------------
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 2.558 s - in reactivewebconsumer.reactivewebconsumer.ReactiveWebConsumerApplicationTests


INITIAL TIME: 2019-05-10T12:53:24.858Z

FINAL TIME: 2019-05-10T12:59:54.453Z

Average 30 secomds


Conclusion:

My personal conclusion, Akka is a more robust framework, because I would choose Akka:

- Best performance

- Higher data load support

- Faster

The Spring Reactor, because I would choose:

- Ease of implementation

- Greater Productivity in Developing

 - Ease of maintenance


Source code here: https://github.com/edsoncezar/spring-akka-reactive

Comentários

Postar um comentário

Postagens mais visitadas deste blog

Simple Application Microservices and MongoDB using Clean Architecture

Imagem
Simple Application Microservices and MongoDB using Clean Architecture Created in mid-2012 by Robert Cecil Martin ("Uncle Bob"), Clean Architecture has as main purposes to be independent of frameworks, easily testable, regardless of the user interface, that is, the user interface can change at will without this reflects in the rest of the system, be independent of database, since it maintains all business rules in the application itself and, finally, be independent of any external agent, and the business rules do not "see" the " outside world ". For start my application I studied some frameworks before choose some, at the end the choose was not for performance, just for an example I choose the fasted way to developed, easy understand, easy codification and quick production instead of best performance. You can ask me, why I would choose facility instead of performance ? First question, do you really need a truck...
Leia mais

Adaline - Adaptive Linear Neuron

Imagem
Just continuing with my explanation about Artificial Neural Networks in a simple way. The Adaline  ( Adaptive Linear  Neuron or later  Adaptive Linear Element )  network, proposed by Widrow and Hoff in 1960, has the same structure as the Perceptron, differentiating only in the algorithm of training. She is an adaptive network, with the Inclusion of a pioneering algorithm for the training of  multi-layered networks, the training algorithm of the Delta rule, the Delta rule algorithm is based on the method of least squares, she has a smoother learning method. Their goal is Perform local iterations to get the function and thus set the values of the weights  when the minimum was found. Through an arbitrary point (vector of weights started randomly), the algorithm runs, at each iteration, the surface of the error function towards the point of Minimum. Adaline is a single layer neural network with multiple nodes where each node accepts multiple...
Leia mais